X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=filterseqscommand.cpp;h=a677eef9729ef5f1035db03f0a5daff9f7280e36;hb=1d898dc6edaf9e9f287fab53bf1f21fb29757a17;hp=3d3062b4cc371848fdbd222b7985415a41e62e57;hpb=fdc1f6eaf544f695fc1511f24bddd7e6069c33ba;p=mothur.git diff --git a/filterseqscommand.cpp b/filterseqscommand.cpp index 3d3062b..a677eef 100644 --- a/filterseqscommand.cpp +++ b/filterseqscommand.cpp @@ -10,25 +10,82 @@ #include "filterseqscommand.h" #include "sequence.hpp" -/**************************************************************************************/ +//********************************************************************************************************************** +vector FilterSeqsCommand::setParameters(){ + try { + CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pfasta); + CommandParameter phard("hard", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(phard); + CommandParameter ptrump("trump", "String", "", "*", "", "", "",false,false); parameters.push_back(ptrump); + CommandParameter psoft("soft", "Number", "", "0", "", "", "",false,false); parameters.push_back(psoft); + CommandParameter pvertical("vertical", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pvertical); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors); + CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir); + CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir); + + vector myArray; + for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); } + return myArray; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "setParameters"); + exit(1); + } +} +//********************************************************************************************************************** +string FilterSeqsCommand::getHelpString(){ + try { + string helpString = ""; + helpString += "The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n"; + helpString += "The filter.seqs command parameters are fasta, trump, soft, hard, processors and vertical. \n"; + helpString += "The fasta parameter is required, unless you have a valid current fasta file. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n"; + helpString += "For example: fasta=abrecovery.fasta-amazon.fasta \n"; + helpString += "The trump option will remove a column if the trump character is found at that position in any sequence of the alignment. Default=*, meaning no trump. \n"; + helpString += "A soft mask removes any column where the dominant base (i.e. A, T, G, C, or U) does not occur in at least a designated percentage of sequences. Default=0.\n"; + helpString += "The hard parameter allows you to enter a file containing the filter you want to use.\n"; + helpString += "The vertical parameter removes columns where all sequences contain a gap character. The default is T.\n"; + helpString += "The processors parameter allows you to specify the number of processors to use. The default is 1.\n"; + helpString += "The filter.seqs command should be in the following format: \n"; + helpString += "filter.seqs(fasta=yourFastaFile, trump=yourTrump) \n"; + helpString += "Example filter.seqs(fasta=abrecovery.fasta, trump=.).\n"; + helpString += "Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n"; + return helpString; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "getHelpString"); + exit(1); + } +} +//********************************************************************************************************************** +FilterSeqsCommand::FilterSeqsCommand(){ + try { + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["fasta"] = tempOutNames; + outputTypes["filter"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "FilterSeqsCommand"); + exit(1); + } +} +/**************************************************************************************/ FilterSeqsCommand::FilterSeqsCommand(string option) { try { - abort = false; + abort = false; calledHelp = false; filterFileName = ""; //allow user to run help - if(option == "help") { help(); abort = true; } + if(option == "help") { help(); abort = true; calledHelp = true; } else { - //valid paramters for this command - string Array[] = {"fasta", "trump", "soft", "hard", "vertical", "outputdir","inputdir", "processors"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser.getParameters(); - ValidParameters validParameter; + ValidParameters validParameter("filter.seqs"); map::iterator it; //check to make sure all parameters are valid for command @@ -36,6 +93,11 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } } + //initialize outputTypes + vector tempOutNames; + outputTypes["fasta"] = tempOutNames; + outputTypes["filter"] = tempOutNames; + //if the user changes the input directory command factory will send this info to us in the output parameter string inputDir = validParameter.validFile(parameters, "inputdir", false); if (inputDir == "not found"){ inputDir = ""; } @@ -44,7 +106,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { it = parameters.find("fasta"); //user has given a template file if(it != parameters.end()){ - path = hasPath(it->second); + path = m->hasPath(it->second); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["fasta"] = inputDir + it->second; } } @@ -52,7 +114,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { it = parameters.find("hard"); //user has given a template file if(it != parameters.end()){ - path = hasPath(it->second); + path = m->hasPath(it->second); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["hard"] = inputDir + it->second; } } @@ -60,28 +122,58 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { //check for required parameters fasta = validParameter.validFile(parameters, "fasta", false); - if (fasta == "not found") { m->mothurOut("fasta is a required parameter for the filter.seqs command."); m->mothurOutEndLine(); abort = true; } + if (fasta == "not found") { + fasta = m->getFastaFile(); + if (fasta != "") { fastafileNames.push_back(fasta); m->mothurOut("Using " + fasta + " as input file for the fasta parameter."); m->mothurOutEndLine(); } + else { m->mothurOut("You have no current fastafile and the fasta parameter is required."); m->mothurOutEndLine(); abort = true; } + } else { - splitAtDash(fasta, fastafileNames); + m->splitAtDash(fasta, fastafileNames); //go through files and make sure they are good, if not, then disregard them for (int i = 0; i < fastafileNames.size(); i++) { if (inputDir != "") { - string path = hasPath(fastafileNames[i]); + string path = m->hasPath(fastafileNames[i]); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { fastafileNames[i] = inputDir + fastafileNames[i]; } } - int ableToOpen; ifstream in; - ableToOpen = openInputFile(fastafileNames[i], in); + int ableToOpen = m->openInputFile(fastafileNames[i], in, "noerror"); + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getDefaultPath() != "") { //default path is set + string tryPath = m->getDefaultPath() + m->getSimpleName(fastafileNames[i]); + m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + fastafileNames[i] = tryPath; + } + } + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getOutputDir() != "") { //default path is set + string tryPath = m->getOutputDir() + m->getSimpleName(fastafileNames[i]); + m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + fastafileNames[i] = tryPath; + } + } + + in.close(); + if (ableToOpen == 1) { - m->mothurOut(fastafileNames[i] + " will be disregarded."); m->mothurOutEndLine(); + m->mothurOut("Unable to open " + fastafileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); //erase from file list fastafileNames.erase(fastafileNames.begin()+i); i--; }else{ - string simpleName = getSimpleName(fastafileNames[i]); + string simpleName = m->getSimpleName(fastafileNames[i]); filterFileName += simpleName.substr(0, simpleName.find_first_of('.')); } in.close(); @@ -95,29 +187,33 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { //if the user changes the output directory command factory will send this info to us in the output parameter outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; - outputDir += hasPath(fastafileNames[0]); //if user entered a file with a path then preserve it + outputDir += m->hasPath(fastafileNames[0]); //if user entered a file with a path then preserve it } } //check for optional parameter and set defaults // ...at some point should added some additional type checking... string temp; + hard = validParameter.validFile(parameters, "hard", true); if (hard == "not found") { hard = ""; } + else if (hard == "not open") { hard = ""; abort = true; } + temp = validParameter.validFile(parameters, "trump", false); if (temp == "not found") { temp = "*"; } trump = temp[0]; temp = validParameter.validFile(parameters, "soft", false); if (temp == "not found") { soft = 0; } else { soft = (float)atoi(temp.c_str()) / 100.0; } - temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; } + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } + m->setProcessors(temp); convert(temp, processors); - hard = validParameter.validFile(parameters, "hard", true); if (hard == "not found") { hard = ""; } - else if (hard == "not open") { abort = true; } - - vertical = validParameter.validFile(parameters, "vertical", false); if (vertical == "not found") { vertical = "T"; } + vertical = validParameter.validFile(parameters, "vertical", false); + if (vertical == "not found") { + if ((hard == "") && (trump == '*') && (soft == 0)) { vertical = "T"; } //you have not given a hard file or set the trump char. + else { vertical = "F"; } + } numSeqs = 0; - } } @@ -126,41 +222,15 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { exit(1); } } - -//********************************************************************************************************************** - -void FilterSeqsCommand::help(){ - try { - - m->mothurOut("The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n"); - m->mothurOut("The filter.seqs command parameters are fasta, trump, soft, hard and vertical. \n"); - m->mothurOut("The fasta parameter is required. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n"); - m->mothurOut("For example: fasta=abrecovery.fasta-amazon.fasta \n"); - m->mothurOut("The trump parameter .... The default is ...\n"); - m->mothurOut("The soft parameter .... The default is ....\n"); - m->mothurOut("The hard parameter .... The default is ....\n"); - m->mothurOut("The vertical parameter .... The default is T.\n"); - m->mothurOut("The filter.seqs command should be in the following format: \n"); - m->mothurOut("filter.seqs(fasta=yourFastaFile, trump=yourTrump, soft=yourSoft, hard=yourHard, vertical=yourVertical) \n"); - m->mothurOut("Example filter.seqs(fasta=abrecovery.fasta, trump=..., soft=..., hard=..., vertical=T).\n"); - m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n\n"); - - } - catch(exception& e) { - m->errorOut(e, "FilterSeqsCommand", "help"); - exit(1); - } -} - /**************************************************************************************/ int FilterSeqsCommand::execute() { try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } ifstream inFASTA; - openInputFile(fastafileNames[0], inFASTA); + m->openInputFile(fastafileNames[0], inFASTA); Sequence testSeq(inFASTA); alignmentLength = testSeq.getAlignLength(); @@ -173,7 +243,7 @@ int FilterSeqsCommand::execute() { m->mothurOutEndLine(); m->mothurOutEndLine(); - if (m->control_pressed) { return 0; } + if (m->control_pressed) { outputTypes.clear(); return 0; } #ifdef USE_MPI int pid; @@ -184,11 +254,15 @@ int FilterSeqsCommand::execute() { ofstream outFilter; - string filterFile = outputDir + filterFileName + ".filter"; - openOutputFile(filterFile, outFilter); + //prevent giantic file name + string filterFile; + if (fastafileNames.size() > 3) { filterFile = outputDir + "merge.filter"; } + else { filterFile = outputDir + filterFileName + ".filter"; } + + m->openOutputFile(filterFile, outFilter); outFilter << filter << endl; outFilter.close(); - outputNames.push_back(filterFile); + outputNames.push_back(filterFile); outputTypes["filter"].push_back(filterFile); #ifdef USE_MPI } @@ -207,7 +281,7 @@ int FilterSeqsCommand::execute() { if(filter[i] == '1'){ filteredLength++; } } - if (m->control_pressed) { for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + if (m->control_pressed) { outputTypes.clear(); for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } m->mothurOutEndLine(); @@ -216,6 +290,12 @@ int FilterSeqsCommand::execute() { m->mothurOut("Length of the original alignment: " + toString(alignmentLength)); m->mothurOutEndLine(); m->mothurOut("Number of sequences used to construct filter: " + toString(numSeqs)); m->mothurOutEndLine(); + //set fasta file as new current fastafile + string current = ""; + itTypes = outputTypes.find("fasta"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setFastaFile(current); } + } m->mothurOutEndLine(); m->mothurOut("Output File Names: "); m->mothurOutEndLine(); @@ -240,46 +320,48 @@ int FilterSeqsCommand::filterSequences() { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); - string filteredFasta = outputDir + getRootName(getSimpleName(fastafileNames[s])) + "filter.fasta"; + string filteredFasta = outputDir + m->getRootName(m->getSimpleName(fastafileNames[s])) + "filter.fasta"; #ifdef USE_MPI - int pid, start, end, numSeqsPerProcessor, num; + int pid, numSeqsPerProcessor, num; int tag = 2001; - vectorMPIPos; + vectorMPIPos; MPI_Status status; MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are MPI_File outMPI; - MPI_File tempMPI; MPI_File inMPI; int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; int inMode=MPI_MODE_RDONLY; - char outFilename[filteredFasta.length()]; + char outFilename[1024]; strcpy(outFilename, filteredFasta.c_str()); - - char inFileName[fastafileNames[s].length()]; + + char inFileName[1024]; strcpy(inFileName, fastafileNames[s].c_str()); MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI); - + if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); return 0; } if (pid == 0) { //you are the root process - MPIPos = setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs + MPIPos = m->setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs numSeqs += num; //send file positions to all processes - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos + for(int i = 1; i < processors; i++) { + MPI_Send(&num, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&MPIPos[0], (num+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } //figure out how many sequences you have to do numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //do your part driverMPIRun(startIndex, numSeqsPerProcessor, inMPI, outMPI, MPIPos); @@ -288,75 +370,70 @@ int FilterSeqsCommand::filterSequences() { //wait on chidren for(int i = 1; i < processors; i++) { - char buf[4]; - MPI_Recv(buf, 4, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status); + char buf[5]; + MPI_Recv(buf, 5, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status); } }else { //you are a child process - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs - numSeqs += num; + MPI_Recv(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); MPIPos.resize(num+1); - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions + numSeqs += num; + MPI_Recv(&MPIPos[0], (num+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); //figure out how many sequences you have to align numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //align your part driverMPIRun(startIndex, numSeqsPerProcessor, inMPI, outMPI, MPIPos); if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); return 0; } - char buf[4]; + char buf[5]; strcpy(buf, "done"); //tell parent you are done. - MPI_Send(buf, 4, MPI_CHAR, 0, tag, MPI_COMM_WORLD); + MPI_Send(buf, 5, MPI_CHAR, 0, tag, MPI_COMM_WORLD); } MPI_File_close(&outMPI); MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else + vector positions = m->divideFile(fastafileNames[s], processors); + + for (int i = 0; i < (positions.size()-1); i++) { + lines.push_back(new linePair(positions[i], positions[(i+1)])); + } #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ - ifstream inFASTA; - openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - - lines.push_back(new linePair(0, numFastaSeqs)); - - driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); + int numFastaSeqs = driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); + numSeqs += numFastaSeqs; }else{ - setLines(fastafileNames[s]); - createProcessesRunFilter(filter, fastafileNames[s]); + int numFastaSeqs = createProcessesRunFilter(filter, fastafileNames[s]); + numSeqs += numFastaSeqs; rename((fastafileNames[s] + toString(processIDS[0]) + ".temp").c_str(), filteredFasta.c_str()); //append fasta files for(int i=1;iappendFiles((fastafileNames[s] + toString(processIDS[i]) + ".temp"), filteredFasta); remove((fastafileNames[s] + toString(processIDS[i]) + ".temp").c_str()); } } if (m->control_pressed) { return 1; } #else - ifstream inFASTA; - openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - - lines.push_back(new linePair(0, numFastaSeqs)); - - driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); + int numFastaSeqs = driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); + numSeqs += numFastaSeqs; if (m->control_pressed) { return 1; } #endif #endif - outputNames.push_back(filteredFasta); + outputNames.push_back(filteredFasta); outputTypes["fasta"].push_back(filteredFasta); } return 0; @@ -368,7 +445,7 @@ int FilterSeqsCommand::filterSequences() { } #ifdef USE_MPI /**************************************************************************************/ -int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector& MPIPos) { +int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector& MPIPos) { try { string outputString = ""; int count = 0; @@ -380,14 +457,15 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil //read next sequence int length = MPIPos[start+i+1] - MPIPos[start+i]; - char buf4[length]; + char* buf4 = new char[length]; MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status); string tempBuf = buf4; if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); } istringstream iss (tempBuf,istringstream::in); + delete buf4; - Sequence seq(iss); gobble(iss); + Sequence seq(iss); m->gobble(iss); if (seq.getName() != "") { string align = seq.getAligned(); @@ -405,11 +483,12 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil if(count % 10 == 0){ //output to file //send results to parent int length = outputString.length(); - char buf[length]; - strcpy(buf, outputString.c_str()); + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status); outputString = ""; + delete buf; } } @@ -420,11 +499,12 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil if(outputString != ""){ //output to file //send results to parent int length = outputString.length(); - char buf[length]; - strcpy(buf, outputString.c_str()); + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status); outputString = ""; + delete buf; } if((num) % 100 != 0){ cout << (num) << endl; m->mothurOutJustToLog(toString(num) + "\n"); } @@ -438,21 +518,24 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil } #endif /**************************************************************************************/ -int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string inputFilename, linePair* line) { +int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string inputFilename, linePair* filePos) { try { ofstream out; - openOutputFile(outputFilename, out); + m->openOutputFile(outputFilename, out); ifstream in; - openInputFile(inputFilename, in); + m->openInputFile(inputFilename, in); - in.seekg(line->start); - - for(int i=0;inum;i++){ + in.seekg(filePos->start); + + bool done = false; + int count = 0; + + while (!done) { if (m->control_pressed) { in.close(); out.close(); return 0; } - Sequence seq(in); + Sequence seq(in); m->gobble(in); if (seq.getName() != "") { string align = seq.getAligned(); string filterSeq = ""; @@ -464,13 +547,27 @@ int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string i } out << '>' << seq.getName() << endl << filterSeq << endl; - } - gobble(in); + count++; + } + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + unsigned long int pos = in.tellg(); + if ((pos == -1) || (pos >= filePos->end)) { break; } + #else + if (in.eof()) { break; } + #endif + + //report progress + if((count) % 100 == 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } } + //report progress + if((count) % 100 != 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } + + out.close(); in.close(); - return 0; + return count; } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "driverRunFilter"); @@ -483,7 +580,7 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { try { #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) int process = 0; - int exitCommand = 1; + int num = 0; processIDS.clear(); //loop through and create all the processes you want @@ -495,18 +592,39 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { process++; }else if (pid == 0){ string filteredFasta = filename + toString(getpid()) + ".temp"; - driverRunFilter(F, filteredFasta, filename, lines[process]); + num = driverRunFilter(F, filteredFasta, filename, lines[process]); + + //pass numSeqs to parent + ofstream out; + string tempFile = filename + toString(getpid()) + ".num.temp"; + m->openOutputFile(tempFile, out); + out << num << endl; + out.close(); + exit(0); - }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } + exit(0); + } } //force parent to wait until all the processes are done for (int i=0;iopenInputFile(tempFile, in); + if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; } + in.close(); remove(tempFile.c_str()); } + - return exitCommand; + return num; #endif } catch(exception& e) { @@ -525,7 +643,7 @@ string FilterSeqsCommand::createFilter() { F.setLength(alignmentLength); - if(soft != 0 || isTrue(vertical)){ + if(trump != '*' || m->isTrue(vertical) || soft != 0){ F.initialize(); } @@ -533,7 +651,7 @@ string FilterSeqsCommand::createFilter() { else { F.setFilter(string(alignmentLength, '1')); } numSeqs = 0; - if(trump != '*' || isTrue(vertical) || soft != 0){ + if(trump != '*' || m->isTrue(vertical) || soft != 0){ for (int s = 0; s < fastafileNames.size(); s++) { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); @@ -541,32 +659,38 @@ string FilterSeqsCommand::createFilter() { #ifdef USE_MPI int pid, numSeqsPerProcessor, num; int tag = 2001; - vector MPIPos; + vector MPIPos; MPI_Status status; MPI_File inMPI; MPI_Comm_size(MPI_COMM_WORLD, &processors); MPI_Comm_rank(MPI_COMM_WORLD, &pid); - char* tempFileName = new char(fastafileNames[s].length()); - tempFileName = &(fastafileNames[s][0]); + //char* tempFileName = new char(fastafileNames[s].length()); + //tempFileName = &(fastafileNames[s][0]); + + char tempFileName[1024]; + strcpy(tempFileName, fastafileNames[s].c_str()); MPI_File_open(MPI_COMM_WORLD, tempFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer if (m->control_pressed) { MPI_File_close(&inMPI); return 0; } if (pid == 0) { //you are the root process - MPIPos = setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs + MPIPos = m->setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs numSeqs += num; //send file positions to all processes - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos + for(int i = 1; i < processors; i++) { + MPI_Send(&num, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&MPIPos[0], (num+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } //figure out how many sequences you have to do numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //do your part MPICreateFilter(startIndex, numSeqsPerProcessor, F, inMPI, MPIPos); @@ -574,16 +698,16 @@ string FilterSeqsCommand::createFilter() { if (m->control_pressed) { MPI_File_close(&inMPI); return 0; } }else { //i am the child process - - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs + MPI_Recv(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); MPIPos.resize(num+1); numSeqs += num; - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions + MPI_Recv(&MPIPos[0], (num+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); //figure out how many sequences you have to align numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //do your part MPICreateFilter(startIndex, numSeqsPerProcessor, F, inMPI, MPIPos); @@ -592,37 +716,27 @@ string FilterSeqsCommand::createFilter() { } MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else + vector positions = m->divideFile(fastafileNames[s], processors); + + for (int i = 0; i < (positions.size()-1); i++) { + lines.push_back(new linePair(positions[i], positions[(i+1)])); + } #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ - ifstream inFASTA; - openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - + int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; - - lines.push_back(new linePair(0, numFastaSeqs)); - - driverCreateFilter(F, fastafileNames[s], lines[0]); }else{ - setLines(fastafileNames[s]); - createProcessesCreateFilter(F, fastafileNames[s]); + int numFastaSeqs = createProcessesCreateFilter(F, fastafileNames[s]); + numSeqs += numFastaSeqs; } if (m->control_pressed) { return filterString; } #else - ifstream inFASTA; - openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - + int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; - - lines.push_back(new linePair(0, numFastaSeqs)); - - driverCreateFilter(F, fastafileNames[s], lines[0]); if (m->control_pressed) { return filterString; } #endif #endif @@ -637,67 +751,81 @@ string FilterSeqsCommand::createFilter() { MPI_Status status; MPI_Comm_rank(MPI_COMM_WORLD, &pid); - if (pid == 0) { //only one process should output the filter - vector temp; temp.resize(alignmentLength+1); - - //get the frequencies from the child processes - for(int i = 0; i < ((processors-1)*5); i++) { - MPI_Recv(&temp[0], (alignmentLength+1), MPI_INT, MPI_ANY_SOURCE, 2001, MPI_COMM_WORLD, &status); - int receiveTag = temp[temp.size()-1]; //child process added a int to the end to indicate what letter count this is for - - if (receiveTag == Atag) { //you are recieveing the A frequencies - for (int k = 0; k < alignmentLength; k++) { F.a[k] += temp[k]; } - }else if (receiveTag == Ttag) { //you are recieveing the T frequencies - for (int k = 0; k < alignmentLength; k++) { F.t[k] += temp[k]; } - }else if (receiveTag == Ctag) { //you are recieveing the C frequencies - for (int k = 0; k < alignmentLength; k++) { F.c[k] += temp[k]; } - }else if (receiveTag == Gtag) { //you are recieveing the G frequencies - for (int k = 0; k < alignmentLength; k++) { F.g[k] += temp[k]; } - }else if (receiveTag == Gaptag) { //you are recieveing the gap frequencies - for (int k = 0; k < alignmentLength; k++) { F.gap[k] += temp[k]; } - } - } - }else{ - - //send my fequency counts - F.a.push_back(Atag); - int ierr = MPI_Send(&(F.a[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); - F.t.push_back(Ttag); - ierr = MPI_Send (&(F.t[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); - F.c.push_back(Ctag); - ierr = MPI_Send(&(F.c[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); - F.g.push_back(Gtag); - ierr = MPI_Send(&(F.g[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); - F.gap.push_back(Gaptag); - ierr = MPI_Send(&(F.gap[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); + if(trump != '*' || m->isTrue(vertical) || soft != 0){ + + if (pid == 0) { //only one process should output the filter + + vector temp; temp.resize(alignmentLength+1); + + //get the frequencies from the child processes + for(int i = 1; i < processors; i++) { + + for (int j = 0; j < 5; j++) { + + MPI_Recv(&temp[0], (alignmentLength+1), MPI_INT, i, 2001, MPI_COMM_WORLD, &status); + int receiveTag = temp[temp.size()-1]; //child process added a int to the end to indicate what letter count this is for + + if (receiveTag == Atag) { //you are recieveing the A frequencies + for (int k = 0; k < alignmentLength; k++) { F.a[k] += temp[k]; } + }else if (receiveTag == Ttag) { //you are recieveing the T frequencies + for (int k = 0; k < alignmentLength; k++) { F.t[k] += temp[k]; } + }else if (receiveTag == Ctag) { //you are recieveing the C frequencies + for (int k = 0; k < alignmentLength; k++) { F.c[k] += temp[k]; } + }else if (receiveTag == Gtag) { //you are recieveing the G frequencies + for (int k = 0; k < alignmentLength; k++) { F.g[k] += temp[k]; } + }else if (receiveTag == Gaptag) { //you are recieveing the gap frequencies + for (int k = 0; k < alignmentLength; k++) { F.gap[k] += temp[k]; } + } + } + } + }else{ + + //send my fequency counts + F.a.push_back(Atag); + int ierr = MPI_Send(&(F.a[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); + F.t.push_back(Ttag); + ierr = MPI_Send (&(F.t[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); + F.c.push_back(Ctag); + ierr = MPI_Send(&(F.c[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); + F.g.push_back(Gtag); + ierr = MPI_Send(&(F.g[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); + F.gap.push_back(Gaptag); + ierr = MPI_Send(&(F.gap[0]), (alignmentLength+1), MPI_INT, 0, 2001, MPI_COMM_WORLD); + } + } + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case + if (pid == 0) { //only one process should output the filter #endif + F.setNumSeqs(numSeqs); - - if(isTrue(vertical) == 1) { F.doVertical(); } + if(m->isTrue(vertical) == 1) { F.doVertical(); } if(soft != 0) { F.doSoft(); } - filterString = F.getFilter(); #ifdef USE_MPI //send filter string to kids - MPI_Bcast(&filterString[0], alignmentLength, MPI_CHAR, 0, MPI_COMM_WORLD); + //for(int i = 1; i < processors; i++) { + // MPI_Send(&filterString[0], alignmentLength, MPI_CHAR, i, 2001, MPI_COMM_WORLD); + //} + MPI_Bcast(&filterString[0], alignmentLength, MPI_CHAR, 0, MPI_COMM_WORLD); }else{ //recieve filterString - char tempBuf[alignmentLength]; + char* tempBuf = new char[alignmentLength]; + //MPI_Recv(&tempBuf[0], alignmentLength, MPI_CHAR, 0, 2001, MPI_COMM_WORLD, &status); MPI_Bcast(tempBuf, alignmentLength, MPI_CHAR, 0, MPI_COMM_WORLD); filterString = tempBuf; if (filterString.length() > alignmentLength) { filterString = filterString.substr(0, alignmentLength); } + delete tempBuf; } MPI_Barrier(MPI_COMM_WORLD); #endif - - + return filterString; } catch(exception& e) { @@ -706,37 +834,46 @@ string FilterSeqsCommand::createFilter() { } } /**************************************************************************************/ -int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* line) { +int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* filePos) { try { ifstream in; - openInputFile(filename, in); + m->openInputFile(filename, in); - in.seekg(line->start); - - for(int i=0;inum;i++){ + in.seekg(filePos->start); + + bool done = false; + int count = 0; + + while (!done) { if (m->control_pressed) { in.close(); return 1; } - Sequence seq(in); + Sequence seq(in); m->gobble(in); if (seq.getName() != "") { if (seq.getAligned().length() != alignmentLength) { m->mothurOut("Sequences are not all the same length, please correct."); m->mothurOutEndLine(); m->control_pressed = true; } - if(trump != '*'){ F.doTrump(seq); } - if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); } + if(trump != '*') { F.doTrump(seq); } + if(m->isTrue(vertical) || soft != 0) { F.getFreqs(seq); } cout.flush(); + count++; } + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + unsigned long int pos = in.tellg(); + if ((pos == -1) || (pos >= filePos->end)) { break; } + #else + if (in.eof()) { break; } + #endif + //report progress - if((i+1) % 100 == 0){ m->mothurOut(toString(i+1)); m->mothurOutEndLine(); } + if((count) % 100 == 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } } - //report progress - if((line->num) % 100 != 0){ m->mothurOut(toString(line->num)); m->mothurOutEndLine(); } - + if((count) % 100 != 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } in.close(); - return 0; + return count; } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "driverCreateFilter"); @@ -745,7 +882,7 @@ int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* } #ifdef USE_MPI /**************************************************************************************/ -int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector& MPIPos) { +int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector& MPIPos) { try { MPI_Status status; @@ -759,19 +896,20 @@ int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& //read next sequence int length = MPIPos[start+i+1] - MPIPos[start+i]; - char buf4[length]; + char* buf4 = new char[length]; MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status); string tempBuf = buf4; if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); } istringstream iss (tempBuf,istringstream::in); - + delete buf4; + Sequence seq(iss); if (seq.getAligned().length() != alignmentLength) { cout << "Alignment length is " << alignmentLength << " and sequence " << seq.getName() << " has length " << seq.getAligned().length() << ", please correct." << endl; exit(1); } if(trump != '*'){ F.doTrump(seq); } - if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); } + if(m->isTrue(vertical) || soft != 0){ F.getFreqs(seq); } cout.flush(); //report progress @@ -794,93 +932,89 @@ int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) { try { #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; - int exitCommand = 1; + int process = 1; + int num = 0; processIDS.clear(); //loop through and create all the processes you want while (process != processors) { - int pid = vfork(); + int pid = fork(); if (pid > 0) { processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later process++; }else if (pid == 0){ - driverCreateFilter(F, filename, lines[process]); + //reset child's filter counts to 0; + F.a.clear(); F.a.resize(alignmentLength, 0); + F.t.clear(); F.t.resize(alignmentLength, 0); + F.g.clear(); F.g.resize(alignmentLength, 0); + F.c.clear(); F.c.resize(alignmentLength, 0); + F.gap.clear(); F.gap.resize(alignmentLength, 0); + + num = driverCreateFilter(F, filename, lines[process]); + + //write out filter counts to file + filename += toString(getpid()) + "filterValues.temp"; + ofstream out; + m->openOutputFile(filename, out); + + out << num << endl; + out << F.getFilter() << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.a[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.t[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.g[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.c[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.gap[k] << '\t'; } out << endl; + + //cout << F.getFilter() << endl; + out.close(); + + exit(0); + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } exit(0); - }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + } } + //parent do your part + num = driverCreateFilter(F, filename, lines[0]); + //force parent to wait until all the processes are done - for (int i=0;ierrorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter"); - exit(1); - } -} -/**************************************************************************************************/ - -int FilterSeqsCommand::setLines(string filename) { - try { - - vector positions; - bufferSizes.clear(); - - ifstream inFASTA; - openInputFile(filename, inFASTA); + //parent reads in and combines Filter info + for (int i = 0; i < processIDS.size(); i++) { + string tempFilename = filename + toString(processIDS[i]) + "filterValues.temp"; + ifstream in; + m->openInputFile(tempFilename, in); - string input; - while(!inFASTA.eof()){ - input = getline(inFASTA); + int temp, tempNum; + string tempFilterString; - if (input.length() != 0) { - if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } - } - } - inFASTA.close(); - - int numFastaSeqs = positions.size(); - - FILE * pFile; - long size; - - //get num bytes in file - pFile = fopen (filename.c_str(),"rb"); - if (pFile==NULL) perror ("Error opening file"); - else{ - fseek (pFile, 0, SEEK_END); - size=ftell (pFile); - fclose (pFile); - } - - numSeqs += numFastaSeqs; - - int numSeqsPerProcessor = numFastaSeqs / processors; - - for (int i = 0; i < processors; i++) { + in >> tempNum; m->gobble(in); num += tempNum; - long int startPos = positions[ i * numSeqsPerProcessor ]; - if(i == processors - 1){ - numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; - bufferSizes.push_back(size - startPos); - }else{ - long int myEnd = positions[ (i+1) * numSeqsPerProcessor ]; - bufferSizes.push_back(myEnd-startPos); - } - lines.push_back(new linePair(startPos, numSeqsPerProcessor)); + in >> tempFilterString; + F.mergeFilter(tempFilterString); + + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.a[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.t[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.g[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.c[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.gap[k] += temp; } m->gobble(in); + + in.close(); + remove(tempFilename.c_str()); } - return 0; + return num; +#endif } catch(exception& e) { - m->errorOut(e, "FilterSeqsCommand", "setLines"); + m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter"); exit(1); } }